Carotenoids are commercially important C-40 isoprenoid compounds used as nutritional supplements, pharmaceuticals and food colorants for humans and as pigments for animal feed. Currently industrially important carotenoids are produced mainly by chemical synthesis (β-carotene, canthaxanthin and astaxanthin) or extraction from natural sources (lutein from marigold, capsanthin from paprika). Production of carotenoids, however, using microorganisms has been achieved in some cases. For example, β-carotene is produced by fermentation with the fungus Blakeslea trispora (Finkelstein et al., U.S. Pat. No. 5,328,845) or by pond culture using the halotolerant alga Dunaliella salina (Borowitzka, J. Biotechnol. 70, 313-321, 1999). Lycopene production has also been reported in B. trispora (Bernasconi et al., International Patent Application Publication No. WO 00/77234). Astaxanthin is produced by fermentation using yeast (Phaffia rhodozyma, (recently renamed Xanthophyllomyces dendorous)) (Jacobson et al., U.S. Pat. No. 6,015,684) or in photobioreactors or open ponds using the alga Haematococcus pluvialis (Lorenz and Cysewski, Trends Biotechnol. 18, 160-167, 1999; Olaizola, J. Appl. Phycol. 12, 499-506, 2000). Such microbial production systems, however, do not produce carotenoids in amounts sufficient for economical industrial scale production.
In the mid-1960's, scientists at Hoffmann-La Roche isolated several marine bacteria that produced the yellow carotenoid zeaxanthin, which has application in poultry pigmentation and in the prevention of age-related macular degeneration in humans. One bacterium, which showed promising levels of zeaxanthin production, was given the strain designation R-1512, and it was deposited at the American Type Culture Collection (ATCC, Manassas, Va., USA) as strain ATCC 21588 (Schocher and Wiss, U.S. Pat. No. 3,891,504). Using the accepted taxonomic standards of that time (classification performed by the Eidg. Technische Hochschule (Zurich) and the National Collection of Industrial Bacteria, Torry Research Station (Aberdeen, Scotland)), the zeaxanthin-producing organism was classified as a member of the genus Flavobacterium, but no species designation was assigned.
An extensive mutagenesis and screening program was subsequently conducted to isolate mutants of R-1512 with higher zeaxanthin productivities. With respect to the presently described work, two such mutants are significant. These mutants, listed in order of their zeaxanthin productivities, are R1534 and R114. A variety of other mutants have been used over the years for biochemical studies of carotenoid biosynthesis (Goodwin, Biochem. Soc. Symp. 35,233-244, 1972; McDermott et al., Biochem. J. 134, 1115-1117, 1973; Britton et al., Arch. Microbiol. 113, 33-37, 1977; Mohanty et al., Helvetica Chimica Acta 83, 2036-2053, 2000).
The early attempts to develop a commercially viable fermentation process for the production of zeaxanthin using classically derived mutants of strain R-1512 were not successful. However, with the advent of molecular biology, the possibility arose that higher zeaxanthin-producing strains could be developed. The first step in this direction was taken with the cloning and sequencing of the carotenoid gene cluster from strain R1534 (Hohmann et al., U.S. Pat. No. 6,087,152 (“Hohmann '152”), which is hereby incorporated by reference as if recited in full herein). Hohmann '152 discloses that the carotenoid genes were functionally expressed in Escherichia coli and Bacillus subtilis resulting in zeaxanthin production in these hosts. Hohmann '152 also disclosed that by modifying the carotenoid gene cluster or by adding a gene from an astaxanthin producing bacterium, it was possible to produce carotenoids other than zeaxanthin (See also Pasamontes et al., EP Application No. 872,554 (“Pasamontes '554”)). Moreover, Pasamontes '554 disclosed that carotenoid production was increased in strain R1534 by introducing cloned carotenoid gene clusters on a multi-copy plasmid.
Despite the enormous structural diversity in isoprenoid compounds, all are biosynthesized from a common C-5 precursor, isopentenyl pyrophosphate (IPP). Up until the early 1990's it was generally accepted that IPP was synthesized in all organisms via the mevalonate pathway (FIG. 1A), even though some experimental results were not consistent with this biogenic scheme (Eisenreich et al., Chemistry and Biology 5, R221-R233, 1998). The discrepancies have since been reconciled by the discovery of an alternate pathway of IPP biosynthesis, the deoxyxylulose (DXP) pathway (Note: The alternate pathway of IPP biosynthesis has been referred to by various names in the scientific literature (DXP pathway, DOXP pathway, MEP pathway, GAP/pyruvate pathway and the non-mevalonate pathway). We use the name DXP pathway here only for the sake of simplicity). The first five reactions of the DXP pathway have been identified (FIG. 1A) (Herz et al., Proc. Nat. Acad. Sci. 97, 2486-2490, 2000), but the subsequent steps leading to formation of IPP have not yet been elucidated.
McDermott et al. (Biochem. J. 134, 1115-1117, 1973) and Britton et al. (J. Chem. Soc. Chem. Comm. p. 27, 1979) showed that crude extracts of zeaxanthin producing mutant strains derived from the original Roche isolates incorporated labeled mevalonate into zeaxanthin. While there was no reason to question this evidence for IPP biosynthesis via the mevalonate pathway, the work was done prior to the discovery of the DXP pathway, and it has been reported that some bacteria (Streptomyces species) possess both pathways for IPP synthesis and that expression of these pathways is temporally regulated (Seto et al., Tetrahedron Lett. 37, 7979-7982, 1996; Dairi et al., Mol. Gen. Genet. 262, 957-964, 2000.). In addition, at present, only a small number of eubacteria have been shown to possess the mevalonate pathway for IPP synthesis. The genes encoding the enzymes of the mevalonate pathway have been cloned and sequenced from some of these bacteria. (Wilding et al., J. Bacteriol. 182, 4319-4327, 2000 and Takagi et al., J. Bacteriol., 182, 4153-4157, 2000).
Several examples exist where the application of metabolic engineering has succeeded in altering or improving carotenoid production in microorganisms (Lagarde et al., Appl. Env. Microbiol. 66, 64-72, 2000.; Wang et al., Biotechnol. Bioeng. 62, 235-241, 1999; Wang et al., Biotechnol. Prog. 16, 922-926, 2000 (and references therein); Sandmann et al., Trends Biotechnol. 17, 233-237, 2000; Misawa and Shimada, J. Biotechnol. 59, 169-181, 1998; Matthews and Wurtzel, Appl. Microbiol. Biotechnol. 53, 396-400, 2000; Albrecht et al., Nature Biotechnol. 18, 843-846, 2000; Schmidt-Dannert et al., Nature Biotechnol. 18, 750-753 2000). For example, E. coli, a non-carotenogenic bacterium, can be engineered to produce carotenoids by introducing the cloned carotenoid (crt) genes from the bacteria Agrobacterium aurantiacum, Erwinia herbicola or Erwinia uredovora (Misawa and Shimada, J. Biotechnol. 59, 169-181, 1998). Harker and Bramley (FEBS Lett. 448, 115-119, 1999) and Matthews and Wurtzel (Appl. Microbiol. Biotechnol. 53, 396-400, 2000) disclosed that carotenoid production in such engineered E. coli strains could be increased by over-expressing the gene coding for 1-deoxy-D-xylulose 5-phosphate synthase (DXPS), the first enzyme in the DXP pathway (E. coli possesses only the DXP pathway for isoprenoid biosynthesis and does not use the mevalonate pathway (Lange et al., Proc. Nat. Acad. Sci. 97, 13172-13177, 2000)). Harker and Bramley (FEBS Lett., 448, 115-119, 1999) also disclosed an increase in the isoprenoid compound ubiquinone-8, in the cells overproducing DXPS. These results supported the hypothesis that limited availability of IPP, resulting from insufficient in vivo activity of DXPS, was limiting the production of carotenoids and other isoprenoid compounds in the engineered strains. Using a similar E. coli system, Kim and Keasling (Biotechnol. Bioeng., 72, 408-415, 2001) disclosed that the combined over-expression of the genes encoding DXPS and the second enzyme of the DXP pathway, DXP reductoisomerase (see FIG. 1A) gave higher carotenoid production than over-expression of just the gene encoding DXPS.
All of these studies were done in E. coli engineered to produce carotenoids. Accordingly, one disadvantage to these studies was that the amount of carotenoids produced by these recombinant E. coli strains were very low compared to the amounts produced by even non-recombinant microorganisms used for industrial production of carotenoids. Furthermore, improved carotenoid production in bacteria by genetic engineering of the IPP biosynthetic pathway has only been shown in organisms that utilize the DXP pathway for IPP formation. No similar studies have been reported for bacteria that produce IPP via the mevalonate pathway.
Metabolic engineering of the mevalonate pathway to improve production of isoprenoid compounds has been reported in yeast. For example, Millis et al. (International Patent Publication No. WO 00/01649) disclosed that production of isoprenoid compounds is increased in Saccharomyces cerevisiae when the gene coding for 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMG-COA reductase, refer to FIG. 1A) is over-expressed. However, it has not been shown that this strategy improves isoprenoid production in bacteria, and in particular, it has not been shown that carotenoid production in bacteria can be improved by amplifying expression of mevalonate pathway genes. While it has been shown that some mevalonate pathway genes from eukaryotes (Campos et al., Biochem. J., 353, 59-67, 2001) and from the bacterium Streptomyces sp. strain CL190 (Takagi et al., J. Bacteriol., 182, 4153-4157, 2000) can be expressed in E. coli, no increase in isoprenoid production was reported in the strains.
In addition to the reactions that form IPP (via the DXP or mevalonate pathways) and the reactions that convert famesyl pyrophosphate (FPP) to various other isoprenoids (e.g., carotenoids, quinones) (FIG. 1B), two other reactions are known to be involved in isoprenoid biosynthesis. IPP isomerase interconverts IPP and its isomer, dimethylallyl pyrophosphate (DMAPP) (FIG. 1B). Two forms of IPP isomerase exist, the type 1 enzyme is well known in eukaryotes and some bacteria, and the newly identified type 2 enzyme that is FMN- and NADP(H)-dependent (Kaneda et al., Proc. Nat. Acad. Sci. 98, 932-937, 2001).
Several reports disclose that in E. coli engineered to produce carotenoids, amplification of native or heterologous type 1 IPP isomerase (idi) genes stimulates carotenoid production (Kajiwara et al., Biochem. J., 324, 421-426, 1997; Verdoes and van Ooyen, Acta Bot. Gallica, 146, 43-53, 1999; Wang et al., Biotechnol. Bioeng. 62, 235-241, 1999). In one report (Wang et al., Biotechnol. Bioeng. 62, 235-241, 1999), it was further disclosed that over-expression of the ispA gene, encoding FPP synthase increased carotenoid production in an engineered carotenogenic strain of E. coli when combined with over-expression of the idi and crtE (GGPP synthase) genes. As is the case for the pathway of IPP biosynthesis, however, it has not been shown that over-expression of genes coding for IPP isomerase or FPP synthase improves carotenoid production in a naturally carotenogenic microorganism. Also, the levels of carotenoids produced in the E. coli strains described above are very low, and it has not been shown that these strategies work in an industrial microorganism where carotenoid production was already high.